Looks like I found simple and elegant solution to DC offset.
I just divided my single heat sink to the two - now both channels are thermally decoupled, and able to reach thermal equilibrium separately.
As Craigtone mentioned above:
Now my amp's channels drifts +-2..2.5mV around zero, with DC servo disconnected.
I ordered a pair of MultiCap PPMFX 8uF 200V, so now I waiting to use a cap's to DC decouple an amp from the source.
I just divided my single heat sink to the two - now both channels are thermally decoupled, and able to reach thermal equilibrium separately.
As Craigtone mentioned above:
Probably this is correct for single amplifier circuit, or it can to work if all transistors, for both channels are the same. But if we have 2 boards with 2 different matched pairs of devices, these will influent the thermal stabilisation, each of another.
Now my amp's channels drifts +-2..2.5mV around zero, with DC servo disconnected.
I ordered a pair of MultiCap PPMFX 8uF 200V, so now I waiting to use a cap's to DC decouple an amp from the source.
Testing with minimal equipment
Problem: how to test an audio amp (Marsh amp) with only basic equipment: signal generator, 50MHz scope, and Fluke DVM?
I've used constant voltage sine wave input, vary the frequency, and measure output voltage to determine frequency response.
I've used square waves at 1KHz and above to measure rise/fall times and look for ringing/overshoot in the 10KHz to 50KHz range.
What other tests can be run with this equipment? Are there simple (or more appropriately minimal component count) circuits that could be built to augment the above? Any easy way to turn a scope (or laptop) into a spectrum analyzer? Being a s/w engr by trade, would a DSP running a FFT algorithm provide a "significant" improvement over my existing equipment?
I can google and find extravagant claims of measuring sub-ppm values using two 10% carbon resistors and some duct tape: I'd like to hear from others what has worked for them, and what has not.
Problem: how to test an audio amp (Marsh amp) with only basic equipment: signal generator, 50MHz scope, and Fluke DVM?
I've used constant voltage sine wave input, vary the frequency, and measure output voltage to determine frequency response.
I've used square waves at 1KHz and above to measure rise/fall times and look for ringing/overshoot in the 10KHz to 50KHz range.
What other tests can be run with this equipment? Are there simple (or more appropriately minimal component count) circuits that could be built to augment the above? Any easy way to turn a scope (or laptop) into a spectrum analyzer? Being a s/w engr by trade, would a DSP running a FFT algorithm provide a "significant" improvement over my existing equipment?
I can google and find extravagant claims of measuring sub-ppm values using two 10% carbon resistors and some duct tape: I'd like to hear from others what has worked for them, and what has not.
Marsh amp is low THD and low noise as well. To perform measurements you need generator and ADC with THD floor 6dB lower then DUT. Marsh amp is 0.003-0.001% THD.
Baisic signal generator has 0.05-0.02% THD at it best, but you need 0.0005% or less.
Buy or build an audio analyser, or good quality ADC module.
http://www.diyaudio.com/forums/equipment-tools/277808-diy-audio-analyzer-ak5397-ak5394a-ak4490.html
http://www.diyaudio.com/forums/equi...onverter-project-audio-measurements-tool.html
PCM4222 Evaluation Module (EVM) - PCM4222EVM - TI Tool Folder
But keep in mind, that dedicated audio analyser has some very important feauters like:
ground decoupling, input overload protection or indication at least...
Since you already have a PC, a $ 149 sound card and free software gets you 90% of an AP ;-)
Jan
Baudline is good for Linux or Mac, it's a free download. baudline signal analyzer - FFT spectrogram
my + for the ARTA too.
But measuring of 0.002% THD can be affected by quite lot of factors: ground loop, PSU noise, EMI/RFI... So it is good idea to get external sound card at least. Its not bad idea to use an optical SPDIF link to connect it to PC. And if the external sound card coming with SMPS type DC adaptor, it is wroth to check with the scope, what's going on that adaptor output.
But measuring of 0.002% THD can be affected by quite lot of factors: ground loop, PSU noise, EMI/RFI... So it is good idea to get external sound card at least. Its not bad idea to use an optical SPDIF link to connect it to PC. And if the external sound card coming with SMPS type DC adaptor, it is wroth to check with the scope, what's going on that adaptor output.
Here is a rerun .... Signal source harmonics (yellow) are 12dB below the HPA harmonics. So I am not sure what the HPA 'produces' itself.
It pretty much has just amplified the noise and harmonics of the generator source.
I'll need to work on an even cleaner signal gen source.
THx-RNMarsh
It pretty much has just amplified the noise and harmonics of the generator source.
I'll need to work on an even cleaner signal gen source.
THx-RNMarsh
Last edited:
What devices did you use?
Wow!
Out of curiosity, what BJT devices are used on the measured board?
Wow!
Out of curiosity, what BJT devices are used on the measured board?
External work well, but I've been using a PCI card and have no problem measuring below 0.001%.
Hello Dr Marsh,
The capacitors that you suggested are arrived today. A pair of MultiCap PPMFX 8uf 200V.
I putted them before 10KOhm Alps RK27 potentiometer. The effect looks good: while amp is DC decoupled, there is no change to the sound - the PPMFX looks to be very high quality, and tonally transparent capacitors, just an excellent suited to decoupling / bypassing. Probably same result expected in passive crossover scenario.
In addition PPMFX cap's are relative inexpensive.
Thank you for you valuable suggestion about those cap's, you proposed a really good solution to my problem, and you pointed me to some fine, reasonably priced cap's for my projects.
Now jus interesting to compare the amp + cap's VS amp + servo using good THD analyser like JensH going to release
The capacitors that you suggested are arrived today. A pair of MultiCap PPMFX 8uf 200V.
I putted them before 10KOhm Alps RK27 potentiometer. The effect looks good: while amp is DC decoupled, there is no change to the sound - the PPMFX looks to be very high quality, and tonally transparent capacitors, just an excellent suited to decoupling / bypassing. Probably same result expected in passive crossover scenario.
In addition PPMFX cap's are relative inexpensive.
Thank you for you valuable suggestion about those cap's, you proposed a really good solution to my problem, and you pointed me to some fine, reasonably priced cap's for my projects.
Now jus interesting to compare the amp + cap's VS amp + servo using good THD analyser like JensH going to release